Functional characterization of cancer- and RASopathies-associated SHP2 and BRAF mutations

(1)

Functional Characterization

of Cancer- and RASopathies-associated SHP2 and BRAF Mutations

Dissertation

zur Erlangung des akademischen Grades doctor rerum naturalium

(Dr. rer. nat.)

im Fach Biologie/ Molekularbiologie

eingereicht an der

Lebenswissenschaftlichen Fakult¨at der Humboldt–Universit¨at zu Berlin

von

M. Sc. Paula Andrea Medina-P´erez

Pr¨asident der Humboldt–Universit¨at zu Berlin Prof. Dr. Jan–Hendrik Olbertz

Dekan der Lebenswissenschaftlichen Fakult¨at Prof. Dr. Richard Lucius

Gutachter: 1. Prof. Dr. Reinhold Sch¨afer 2. Prof. Dr. Hanspeter Herzel 3. Prof. Dr. Holger S¨ultmann

Tag der m¨undlichen Pr¨ufung: 12.03.2015

(2)

Abstract

Deregulation of the Ras/MAPK signaling is implicated in a wide variety of human diseases, including developmental disorders and cancer. In the last years, a group of developmental disorders, characterized by an overlapping phenotype in patients, was clustered under the term RASopathies. These disorders result from germline mutations in genes encoding key components of the Ras/MAPK signaling cascade. Although the incidence of solid tumors in patients suﬀering from these disorders is rather low, reports on diﬀerent forms of leukemia have considerably increased.

In this work, a group of mutations in the genes SHP2/PTPN11 and BRAF, both key regulators of the MAPK signaling pathway and implicated in RASopathies and cancer, were selected for expression in well-established cell systems for a comprehensive molecular and phenotypic characterization using high-throughput approaches and functional assays.

Synthetic cDNA sequences carrying the SHP2 mutations T42A, E76D, I282V (Noonan syndrome-associated), E76G, E76K, E139D (Noonan- and leukemia-associated), T468M (LEOPARD syndrome -associated) and the BRAF mutations Q257R, S467A, L485F and K499E (cardio-facio-cutaneous syndrome-associated) were shuttled into the modified lentiviral vector pCDH-EF1-IRES-GFP. The non-tumorigenic human cell lines MCF10A, BJ-ELB and HA1EB and the rat preneoplastic 208F fibroblasts were transduced with recombinant lentiviral particles carrying either SHP2 or BRAF mutations to identify their potential roles in neoplastic transformation. MCF10A and BJ-ELB cells overexpressing SHP2 mutations displayed a growth arrest morphology, while BRAF mutations induced cell proliferation and a transformation phenotype. In contrast, both SHP2 and BRAF mutations promoted a spindle-like cell morphology, cell proliferation, density- and anchorage-independent growth in 208F rat fibroblasts. These results suggested that RASopathies-associated mutations in SHP2 and BRAF confer a transformation phenotype in vitrosimilar to the classical H-Ras and BRAF oncogenes. To further investigate whether mutations in SHP2 contribute to tumor growthin vivo, 208F cells expressing either SHP2 wild-type, E76G or T468M mutations were subcutaneously injected in nude mice. Interestingly, cells harboring mutations on SHP2, as well as overexpressing wild-type SHP2, promoted tumor growth.

Reverse-phase protein array (RPPA) and immunoblot assays revealed that RASopathies- associated mutant SHP2 and BRAF proteins constitutively activate the Ras/MAPK signaling pathway in a moderate manner compared to the oncogenic BRAF V600E. Furthermore, to identify modifications in the protein interaction mechanisms of SHP2 mutant proteins, tandem affinity purification (TAP) and yeast-two-hybrid assays were performed using the isogenic dox-inducible HEK-TREx cell system. E76G and T42A SHP2 mutant proteins showed an increased binding strength to GAB1 compared to the wild-type protein. Fi- nally, to investigate the impact of these mutations on gene transcription, a microarray analysis of mRNA from HEK-TREx cells expressing mutant transgenes was conducted. A gene cluster was found to be commonly regulated in both RASopathies-associated BRAF and the oncogenic V600E mutation. This is the first report on transcriptome analysis of RASopathies-associated mutations.

The findings of this study might be useful for a better understanding of the downstream mechanisms of RASopathies-related signaling pathways and their involvement in cancer progression. Moreover, new candidate therapeutic targets for the effective treatment of patients suffering from Ras/MAPK pathway-associated developmental disorders could be evaluated in the future.

(3)

Acknowledgments

This work could not be possible without the encouragement and collaboration of many people. I am especially grateful to my supervisor Prof. Reinhold Sch¨afer for his guidance and valuable advices. I also want to thank Prof. Christine Sers, for the constructive discussions on science and the struggles of the scientiﬁc career.

I want to thank the collaborators of the MUTANOM Consortium, specially to Prof.

Holger Sültmann and Dr. Julia Starmann from the German Cancer Research Center (DKFZ) for the RPPA analysis, Dr. Gerard Joberty from Cellzome, Dr. Artur Muryadan and PD. Dr. Bodo Lange from the Max Planck Institute for Molecular Genetics for the TAP purification and mass spectrometry and Dr. Sean-Patrick Riechers from the Max-Delbrück Center (MDC) for the yeast-two-hybrid assay. Thanks to Maria Stecklum and Dr. Jens Hoffmann from the Experimental Pharmacology and Oncology (EPO) GmbH, who provided thein vivo experiments in nude mice.

I am thankful to Bertram Klinger and Dr. Karsten J¨urchott for the valuable discussions and clear explanations on bioinformatics and the microarray analysis.

I also would like to thank Sabine Bobbe, Conny Gieseler and Kerstin M¨ohr for their technical help in minipreps, platting cells and southern blot preparations. A special thank to Jana Keil, who had always time for explaining new methods and answering all type of technical and do-you-know-where-is-it questions.

I am very greatul to all members of Molecular Tumor Pathology Lab, for their helpful comments and suggestions on my results. Thanks to Stephanie Seibt, Shila Mang-Fatehi, Sha Liu, Natalia Kuhn, Christina Kuznia, Dirk Schumacher, Felix Bormann and other members of the lab for the nice working atmosphere and support. Thanks to Dr. Torben Redmer for the german style corrections.

A big thank to my family and friends, particularly my mother, who despite the physical distance was always present and motivate me not to give up, and my sister, who cheered me up in the diﬃcult moments.

Finally, my deeply gratitude goes to my Husband Manuel G´ongora, for his constant support, love, motivation and inﬁnite patience during this time. Also to our Son Miguel Angel, who confront me with my ability to explain him cancer biology in the most simple way.

(4)

2.2.1 SHP2^wt/mutantsdo not affect cell morphology of the epithelial HA1EB cells but BRAF^wt/mutants influence cell growth pattern . 22 2.2.2 SHP2, but not BRAF, decelerates cell growth of BJELB fibroblasts 23 2.2.3 The MCF10A epithelial cells exhibited a senescence-like state with SHP2^wt/mutants and a transformed phenotype with BRAF^mutants 25 2.2.4 SHP2 Mutations confer a transformed phenotype in rat fibroblasts 25 2.2.5 CFC-associated BRAF and SHP2 mutations stimulate cell proliferation in rat fibroblasts . . . 29

2.2.6 Mutations in SHP2 and BRAF promote anchorage - independent growth of 208F cells . . . 31

2.2.7 NS- and LS-associated SHP2 mutations promote tumor growth in nude mice . . . 32

2.3 Eﬀects on the MAPK signaling cascade . . . 35

2.3.1 Signaling studies in isogenic HEK-TREx cells by Reverse Phase Protein Array (RPPA) . . . 35

2.3.2 SHP2 and BRAF mutations eﬀects on signaling in 208F cells . . 40

2.4 Eﬀects of mutations in SHP2 on Protein-protein interactions . . . 41

2.4.1 Yeast Two-Hybrid Assay . . . 41

2.4.2 Tandem Aﬃnity Puriﬁcation assay . . . 45

2.4.2.1 Validation of the SHP2^mutants-GAB1 complex. . . 47

2.5 Gene regulation in SHP2 and BRAF mutants at the transcription level . 50 2.5.1 Overlapping gene sets within BRAF- and SHP2-HEK-TREx . . 50

2.5.2 Overlapping gene sets within CFC-associated mutants and oncogenic BRAF . . . 52

3 DISCUSSION 54 3.1 Phenotype comparison of SHP2 in human cell lines . . . 54

3.2 Analysis of the eﬀects of NS/LS- and leukemia-associated SHP2 mutations in rat ﬁbroblasts . . . 56

3.2.1 Eﬀects of SHP2 mutations on signal trasduction . . . 57

3.3 CFC-associated BRAF mutations confer a transformed phenotype in preneoplastic rat ﬁbroblasts . . . 59

3.3.1 MAPK and AKT signaling impairment in CFC-associated BRAF- expressing cells . . . 59

3.4 Modiﬁed protein interactions in NS- and cancer- associated SHP2 mutations 61 3.5 Microarray analysis . . . 62

3.5.1 Eﬀects of cancer- and NS/LS-associated SHP2 mutations on gene transcription . . . 62

3.5.2 Eﬀects of CFC-associated BRAF mutations on gene transcription 62 3.6 Outlook . . . 63

(6)

4 MATERIALS AND METHODS 65

4.1 Materials . . . 65

4.1.1 Chemicals . . . 65

4.1.2 Cell culture reagents . . . 66

4.1.3 Restriction enzymes . . . 67

4.1.4 Consumables . . . 67

4.1.5 Commercial kits . . . 68

4.1.6 Antibodies . . . 68

4.1.7 Buﬀers and media . . . 69

4.1.8 Vector backbones . . . 70

4.1.9 Competent bacteria strains . . . 71

4.1.10 Cell lines . . . 71

4.1.11 Software . . . 72

4.1.12 Lab equipment . . . 72

4.1.13 Company register . . . 73

4.2 Molecular biology methods . . . 74

4.2.1 Synthesis of wild-type and mutated genes . . . 74

4.2.2 Gateway^® Cloning . . . 74

4.2.3 Generation of the EF1α promoter-driven pLenti6 expression vector 75 4.2.4 Generation of the new lentiviral expression vector pCDH-EF1a-Puro 75 4.2.5 Transformation of plasmid DNA in competent cells . . . 76

4.2.6 Plasmid DNA puriﬁcation from transformed bacteria . . . 76

4.2.7 Agarose gel electrophoresis . . . 76

4.2.8 RNA Isolation . . . 76

4.3 Cell biology methods . . . 77

4.3.1 Culture of mammalian cell lines . . . 77

4.3.2 Thawing of cell lines . . . 77

4.3.3 Cryopreservation of cell lines . . . 77

4.3.4 Proliferation assay . . . 77

4.3.5 Soft agar assay . . . 78

4.3.6 Trasient tansfection of cells . . . 78

4.3.7 Production of lentiviral particles . . . 79

4.3.8 Lentiviral transduction . . . 79

4.3.9 Generation of stable transduced cell populations . . . 79

4.3.10 Generation of stable dox-inducible T-REx-HEK293 isogenic cell lines 80 4.4 Protein biochemistry methods . . . 80

4.4.1 Whole cell protein extraction . . . 80

4.4.2 Determination of protein concentration . . . 81

4.4.3 SDS-polyacrylamid gel electrophoresis . . . 81

4.4.4 Western blotting . . . 81

4.4.5 Reverse-phase protein array . . . 82

4.4.6 Tandem aﬃnity puriﬁcation . . . 82

4.4.7 Yeast two-hybrid system assay . . . 83

4.4.8 Co-immunoprecipitation . . . 84

4.5 Phenotypic characterization methods . . . 84

4.5.1 Xenotransplantation . . . 84

4.5.2 Immunohistochemistry . . . 84

4.6 Bioinformatic analysis . . . 85

(7)

4.6.1.1 Data pre-processing . . . 85 4.6.1.2 Gene Ontology and pathway analysis . . . 86

APPENDIX XII

A.1 RPPA assays . . . XII A.2 Y2H assay . . . XVIII A.3 Tandem aﬃnity puriﬁcation . . . XXI A.4 Microarray analysis . . . XXII

BIBLIOGRAPHY XXII

ABBREVIATIONS XL

SELBSTST ¨ANDIGKEITSERKL ¨ARUNG XLIV

(8)

List of Figures

1.1 The Ras/MAPK signaling pathway. . . 2 1.2 Activation of the protein tyrosine phosphatase SHP2. . . 3 1.3 Regulation of Ras/MAPK pathway by SHP2 and genes aﬀected in devel-

opmental disorders. . . 6 1.4 Mutations associated with Noonan syndrome and leukemia. . . 7 2.1 SHP2 and BRAF protein domains with the localization of leukemogenic,

Noonan/LEOPARD and CFC mutations. . . 15 2.2 Transfection efficiency of target cell lines . . . 17 2.3 EF1α promoter cloning in pLenti6 expression vector . . . 18 2.4 Comparison of lentiviral vectors in infection efficiency in 208F and Cos7 19 2.5 Generation of the expression vector pCDH-gate-Puro . . . 20 2.6 Optimization of H-Ras ectopic expression in 208F rat fibroblasts . . . 21 2.7 Morphological changes of HA1EB after transduction with SHP2 or BRAF 23 2.8 Cell morphology of BJELB after transduction with SHP2 or BRAF . . . 24 2.9 Overexpression of mutant SHP2/BRAF reduce cell proliferation in BJ-ELB

cells . . . 24 2.10 Morphological phenotype of MCF10A after transduction with SHP2 or

BRAF . . . 26 2.11 CFC-associated BRAF mutations lead to moderate ERK1/2 activation in

mammary epithelial cells . . . 26 2.12 Cell density-independent growth of 208F after transduction with SHP2 . 28 2.13 Morphological changes of 208F after transduction with BRAF . . . 29 2.14 Cell length eﬀect of SHP2 and BRAF mutations on 208F . . . 29 2.15 SHP2/BRAF mutants inﬂuences density-dependent cell proliferation . . 30 2.16 SHP2 and BRAF mutations promote anchorage-independent growth of

208F cells . . . 32 2.17 SHP2 mutations promote growth of solid tumors in nude mice . . . 33 2.18 MEK/ERK activation in xenografts of rat ﬁbroblasts carrying NS/LS-

associated SHP2 mutants . . . 34 2.19 Induction of YFP-SHP2 overexpression in HEK-TREx cell lines . . . 35 2.20 Induction of YFP-BRAF overexpression in HEK-TREx cell lines . . . . 36 2.21 RPPA analysis of SHP2 mutations in isogenic HEK-TEx cell lines . . . 38 2.22 RPPA analysis of BRAF mutations in isogenic HEK-TEx cell lines . . . 39 2.23 MAPK signaling after overexpression of SHP2 mutants . . . 40 2.24 MAPK signaling after overexpression of BRAF mutants . . . 41 2.25 Distribution of protein interaction partners of SHP2 . . . 42 2.26 Overexpression of SHP-2-TRex cells after induction with doxycycline . . 45 2.27 Tandem Aﬃnity Puriﬁcation (TAP) assay revealed an increased binding

of SHP2 mutants to GAB1 . . . 46 2.28 Stable SHP2-overexpressing HEK293 cells . . . 47

(9)

2.29 SHP2-GAB1 interaction complex in HEK293-SHP2 cells . . . 48 2.30 SHP2-GAB1 interaction complex in HEK-TREx-YFP cells after EGF

treatment . . . 48 2.31 Common targets in SHP2- and BRAF-TREx expression proﬁles . . . 51 2.32 GO analysis and heatmap of commonly regulated genes in CFC-associated

and V600E BRAF HEK-TREx cells. . . 52 A.1 RPPA analysis of AKT, STAT3 and GSK3α β in isogenic SHP2-TREx cellsXII A.2 RPPA analysis of MEK1/2, ERK1/2, Ras and cyclin D1 in isogenic

SHP2-TREx cells . . . XIII A.3 RPPA analysis of PI3K-p85α/110α β, mTOR, and p70S6K in isogenic

SHP2-TREx cells . . . XIV A.4 RPPA analysis of AKT, STAT3 and GSK3α βin isogenic BRAF-TREx cellsXV A.5 RPPA analysis of MEK1/2, ERK1/2 and mTOR in isogenic BRAF-TREx

cells . . . XVI A.6 RPPA analysis of p70S6K and PI3K-p85α/110αβin isogenic BRAF-TREx

cells . . . XVII

(10)

List of Tables

1.1 Cancer mutations selected for analysis in the MUTANOM consortium . 9 2.1 Missense mutations in PTPN11 and BRAF selected for functional studies. 14 2.2 Protein targets selected for RPPA analysis . . . 37 2.3 High confidence preys from the yeast two-hybrid assay with SHP2^wt as bait. 44 2.4 Number of significant regulated genes in SHP2- and BRAF-HEK-TRex cells 50 2.5 Signaling pathways affected by CFC- and cancer-associated BRAF mutations 53 4.9 Assays performed with T-REx-HEK293 cells . . . 80 A.10 Literature search for SHP2 protein interaction partners. . . XVIII A.11 Preys obtained after yeast-two-hybrid with SHP2 wild-type as bait . . . XX A.12 Tandem affinity purification assay of SHP2-HEK-TRex cells. . . XXI A.13 Significant regulated genes in NS/LS-associated SHP2 mutants . . . XXII A.14 Significant regulated genes in cancer- and CFC-associated BRAF mutants XXIII A.15 Overlapping regulated genes between NS/LS-associated SHP2 mutations

and BRAF^V⁶⁰⁰^E . . . XXIV A.16 GO analysis and heatmap of commonly regulated genes in CFC-associated

and V600E BRAF HEK-TREx cells. . . XXV

(11)

1 INTRODUCTION

1.1 The Ras/MAPK signaling pathway

The Ras/mitogen-activated protein kinases (MAPK) pathway is among the most wide- ranging regulatory mechanisms of signal transduction in the eukaryotic organisms.

The MAPK pathway has been associated with diverse biological processes varying from development, cell growth, proliferation, diﬀerentiation, migration and apoptosis.

Therefore, it is not surprising that deregulation of the MAPK pathway plays also a central role in many cancer types, not only because its extensive capability to crosstalk with other signaling pathways, but also because many of its core components are encoded by oncogenes frequently found mutated, including the small GTPase Ras and RAF proteins and the receptor tyrosine kinase epidermal growth factor receptor (EGFR).

This signaling cascade becomes activated when extracellular growth factors or cytokines bind to the corresponding receptors, commonly receptor tyrosine kinases (RTK). (ﬁg.

1.1). After ligand binding, activation of the receptor leads to the recruitment of adaptor proteins to the cytosolic membrane, which in turn, transduce the extracellular signal stimulus to intracellular components (Lemmon and Schlessinger, 2010).

1.1.1 Post-translational regulation of the MAPK signaling pathway by protein tyrosine phosphatases

There are a variety of molecular mechanisms that regulate the activation of core components of the MAPK pathway. Ras proteins, for example, undergo palmitoylation and farnesylation to enable membrane association under normal physiological conditions (for review, see Roberts and Der, 2007). Also SUMOylation of MEK1 and MEK2 pro-

teins has been reported to act as a negative modulation mechanism to downregulate the MAPK pathway (Kubota et al., 2011). Nevertheless, phosphorylation of protein kinases is the most well elucidated post-translational modiﬁcation (PTM) that acts as a positive regulator of this signaling cascade. Equivalently, dephosphorylation by protein phosphatases has gained strength as a key regulation event that inﬂuences not only protein activation, but also localization and stability.

Protein phosphatases are derived from different ancestors and are classified in mainly two groups: protein serine/threonine phosphatases and protein tyrosine phosphatases (PTPs). To date, 107 members of the human PTP superfamily have been identified to share the common cysteine-dependent signature motif HC(X)5R to remove a phosphate group from the substrate (Tonks, 2006). According to its structural homology and

(12)

SHP2

RTK

Ras-GTP

Raf

MEK ERK SHC GRB2 ^RasGEF

Cell proliferation and differentiation

Ras-GDP

RasGAP RasGEF

Cytoplasm Nucleus Transcription

factor activation GAB1

Figure 1.1: The Ras/MAPK signaling pathway. Extracellular growth factors or cytokines binds to receptor tyrosine kinases (RTKs), leading to its dimerization and cross-phosphorylation of tyrosine residues. Phosphorylation of the RTK leads to the recruitment of adaptor proteins such growth-factor receptor bound protein 2 (GRB2), which binds to the guanine nucleotide exchange factor (RasGEF) Son Of Sevenless (SOS). SOS activates Ras by binding to Ras-GDP complexes and promotes the exchange of GDP (guanosine diphosphate) to GTP (guanosine triphosphate). Next, Activated Ras (Ras-GTP) binds the serine/threonine kinase RAF, which then activates a phosphorylation cascade of MEK and ERK proteins that are translocated to the nucleus and activate transcription factors. Modiﬁed from Lemmon and Schlessinger (2010); Ahearn et al. (2012).

substrate specificity, the PTP family is subdivided in four classes: phosphotyrosine-specific phosphatases, dual-specificity phosphatases, cdc25 phosphatases and low molecular PTPs.

1.1.2 The protein tyrosine phosphatase SHP2

Mammalian SHP2 (also known as SH-PTP2, SH-PTP3, PTP2C, PTP1D and Syp) is an ubiquitously expressed non-transmembrane protein-tyrosine phosphatase that belongs to the phosphotyrosine-speciﬁc phosphatases and is encoded by the gene PTPN11 in the human chromosome 12q24. It shares homologues in Drosophila (Corkscrew) and C. elegans (Ptp2). SHP2 was identiﬁed by R. M. Freeman et al. (1992), shortly after

corkscrew (csw) (Perkins et al., 1992).

1.1.2.1 SHP2 protein activation

SHP2 contains two tandemly arranged src-homology 2 region domains (SH2 domains), followed by a catalytic phosphatase domain (PTP domain), two tyrosine residues at the C-terminus and a proline-rich sequence. The crystal structure revealed an autoinhibitory mechanism of the catalytic site that regulates its basal state (Hof et al., 1998). SHP2 activity is suppressed by intramolecular interactions between residues in the backside loop of the N-terminal SH2 domain (N-SH2) and the catalytic surface of the PTP domain (ﬁg.

(13)

to phosphorylated tyrosine residues on RTKs, cytokine receptors, and/or scaﬀolding adaptors, such as insulin receptor substrate, ﬁbroblast growth factor receptor substrate, or GRB2-associated binding (GAB) proteins. Phosphotyrosyl (pY) peptide binding to the N-SH2 domain disrupts the autoinhibitory interaction leading to an equilibrium shift and hence to an open conformation of the PTP domain and its catalytic activation (Ahmad et al., 1993; R. M. Freeman et al., 1992).

C C-SH2

PTP Y₅₄₂ Y₅₈₀ N-SH2

N-SH2 C-SH2 PTP

Binding protein Substrate

P P

Y Y Y_P

Inactive

Active

P P

Figure 1.2: Activation of the protein tyrosine phosphatase SHP2. SHP2 is conformed by two src- homology 2 domains (N-SH2 and C-SH2), a protein tyrosine phosphatase domain (PTP) and two C-tail tyrosine residues (Y542 and Y580). SHP2 is kept in a closed conformation by the interaction of N-SH2 and the PTP domains that blocks the catalytic site. Upon binding of phosphotyrosine proteins (pY) to SH2 domains, the PTP domain is exposed, so substrates can bind the active site. Adapted from Grossmann et al. (2010); Qiu et al. (2014).

Bennett et al. (1994) identiﬁed SHP2 as a positive regulator of the platelet-derived growth factor receptorβ (PDGFR) by binding GRB2 and the PDGF receptor directly.

Additionally, SHP2 has been found to act as a modulator upstream of the Ras/MAPK signaling cascade by regulating Sprouty activity through tyrosine dephosphorylation. This results in dissociation of Sprouty proteins from GRB2, enabling the positive regulation of ERK activation (Hanafusa et al., 2004; Jarvis et al., 2006). Additionally, SHP2 binding to c-Met-activated GAB1 leads to c-Met speciﬁc signaling activation (Schaeper et al., 2000)

1.1.2.2 Biological relevance of SHP2

Due to its ubiquitous expression and its cell-type speciﬁc signaling outcome, SHP2 acts as a positive regulator in many signaling cascades that includes the Jak/STAT, the NFkB and the Ras/MAPK pathway (Grossmann et al., 2010). Therefore, it is not surprising that SHP2 plays a central role in a broad spectrum of cellular processes such as cell proliferation, diﬀerentiation and embryonic development. For example, Saxton et al.

(1997) demonstrated that Shp2 is essential during gastrulation in the organization of axial mesoderm. They generated a mouse model by introducing an internal deletion of residues 46-110 in the N-terminal SH2-domain and found that mice homozygous for the mutant allele diedin utero at mid-gestation. The mutant embryos showed uncompleted

(14)

turning, recognized by the disorganized neuroectoderm, and perturbed development of the vascular system. These observations were later confirmed in tissue-specific null mutations by different groups (Kontaridis et al., 2008; Princen et al., 2009).

Additionally, SHP2 is required for branching morphogenesis of the kidney in mammals (Schaeper et al., 2000; Willecke et al., 2011) and the development and maintenance of the nervous system (Grossmann et al., 2010). It also regulates cell fate during cardiomyogenesis and angiogenesis (Mannell et al., 2008). Recently, SHP2 has been also found to inﬂuence the diﬀerentiation of goblet and paneth cells in the murine intestine by controlling of the canonical Wnt/β-catenin signaling pathway (Heuberger et al., 2014).

1.1.2.3 Role of SHP2 in cancer

According to the Catalogue of Somatic Mutations in Cancer (COSMIC), there are 142 mutations in the PTPN11 gene associated with cancer. Most of them are located in the N-SH2 domain, followed by mutations encoding for the terminal tail of the PTP domain. These somatic mutations are frequently associated with hematopoietic malignancies, from which approximately 35% are related with juvenile myelomonocytic leukemia (JMML) and in a lower incidence, with acute myeloid leukemia (AML), chronic myelomonocytic leukemia (CMML), myelodysplastic syndrome (MDS), and B-acute lymphoblastic leukemia (B-ALL) (Forbes et al., 2008; Grossmann et al., 2010).

To date, 65 substitution mutations in 21 amino acid positions of the N-SH2 domain have been reported, being the position Glu76 the most frequently mutated aminoacid. The E76K mutation alone has been reported in 85 cancer samples, most of them related with diﬀerent forms of leukemia, but also in single cases of lung and colon cancer. Two additional amino acids, Ala72 (A72) and Asp61 (D61), are also frequently mutated.

The PTP domain has as well a relative high incidence of mutations (70 substitution mutations in 38 positions). Only the position Gly503 (G503) accounts for 36 reported cancer samples closely followed by Ser502 (19 counts). On the other hand, the C-SH2 domain shows a relative low mutation frequency (16 substitution mutations) compared to the N-SH2 domain, being mutations in the position Glu139 (E139) the most common.

SHP2 and H. pylori-CagA

Epidemiological studies have demonstrated that there is a strong correlation between increased SHP2 protein expression and gastric carcinogenesis in patients infected with Helicobacter pylori CagA-positive strains. H. pylori CagA is a 120-145 KDa protein and a tyrosine-phosphorylated by Src family protein-tyrosine kinase. CagA was ﬁrst identiﬁed as a virulence factor ofH. pylori (CagA positive strains) and associated with peptic ulcers.

Additionally, infections with cagA-positive H. pylori strains are strongly associated with gastric adenocarcinoma (Hatakeyama, 2006b; Kim et al., 2010; Jiang et al., 2013). Not only the increased expression of SHP2 but also the activation of the IL6/gp130/STAT3 signaling pathway has been shown to be implicated in the development of gastric cancer (Lee et al., 2010). The interaction of the phosphorylated CagA protein with SHP2

(15)

dephosphorylation, promoting an elongated host-cell shape termed the ”hummingbird phenotype” and increased cell motility (Higashi et al., 2002; Tsutsumi et al., 2006).

Moreover, the eﬀector protein CagA may also modulate a similar cell motility response by targeting the c-Met receptor, which in turn, recruits the adapter proteins GAB1 and SHP2 in epithelial cells (Churin et al., 2003).

1.1.3 Post-translational regulation of the MAPK signaling pathway by the protein kinase BRAF

BRAF, encoded in the human chromosome 7q34, is a serine/threonine protein kinase that belongs to the RAF family, key regulators of the MAPK pathway. Somatic mutations in this gene are associated with 8% of all human cancers, including colorectal cancer, malignant melanoma, thyroid carcinoma and non-Hodgkin lymphoma (Davies et al., 2002).

Recent studies associate germline mutations in BRAF with developmental disorders, such as the cardio-facio-cutaneous syndrome (CFC syndrome).

BRAF protein consists of three conserved regions, which share the following domains with other RAF proteins: two regulatory CR1 and CR2 domains and a CR3 region that contains a motif called the negative-charge regulatory region (N-region), a glycine-rich loop, a catalytic loop and the activation domain or kinase domain (Sithanandam et al., 1990; Wellbrock and Marais, 2005). The most frequently somatic point mutation found in cancer is V600E (>90%) is located in the activation segment of the kinase domain (CR3). As a consequence, the V600E mutation derives in a BRAF protein with an

elevated kinase activity.

BRAF Activation

There are three Raf paralogs in humans (A-Raf, B-Raf and C-Raf) coding for Raf proteins that are activated after extracellular stimuli and by binding of Ras-GTP to the cysteine-rich domain located in the CR1 region (ﬁg. 2.1).

BRAF is a protein kinase that catalizes the phosphorylation of serine and threonine residues in consensus sequences of protein substrates using ATP. The products of this reaction are ADP and a phosphorylated protein. Under normal conditions, BRAF is kept inactive by auto-inhibition of the Ras-GTP-binding CR1 domain and the hinge domain. In contrast, oncogenic BRAF is constitutively active independently of mitogenic activation.

BRAF plays an important role in endothelial development. This feature was demonstrated by Wojnowski et al. (1997), who developed a mouse with a targeted disruption in the Braf gene. Heterozygous mice did not exhibited obvious defects. However, homozygous Braf-deﬁcient mice showed an increased number of endothelial precursor cells, enlarged blood vessels and died of vascular defects during midgestation.

In 2007, Schubbert et al. (2007) reviewed the developmental disorders associated with mutations in the Ras/MAPK signaling pathway. In the CFC syndrome, there was neither an overlap in the mutation pattern, nor an association with cancer, quite opposite from the other developmental syndromes. Nevertheless, another study identiﬁed mutations in

(16)

diﬀerent components of the MAPK cascade, including BRAF, that were implicated in acute lymphoblastic leukemia (ALL) and non-Hodgkin lymphoma (Aoki and Matsubara, 2013).

1.2 Ras/MAPK pathway deregulation in developmental disorders

The term “RASopathies”was coined in an attempt to cluster developmental disorders that result from germline mutations in genes encoding components of the canonical Ras/MAPK signaling pathway. Some affected genes includePTPN11, NRAS,HRAS, BRAF, RAF1, SOS, MEK1, MEK2 (fig. 1.3), and share phenotypic features that includes craniofacial manifestations, cardiac, skin, muscular and ocular abnormalities, neurocognitive disabilities and an increased risk of developing cancer (Rauen et al., 2011). Some of these disorders, such as the cadio-facio-cutaneous syndrome (CFC), Noonan, LEOPARD (acronym for multiple Lentigines, Electrocardiographic conduction abnormalities, Ocular hypertelorism, Pulmonic stenosis, Abnormal genitalia, Retardation of growth, and sensorineural Deafness) and Costello syndromes are difficult to diagnose due to overlapping symptoms.

SHP2

RTK

Ras-GTP

Raf

MEK

ERK SHC GRB2 SOS1

Cell growth and differentiation Ras-GDP

NF1

HRAS: CS KRAS: NS, CFCS

RAF1: NS, LS BRAF: NS, LS, CFCS

MEK1: NS, CFCS MEK2: CFCS PTPN11: NS, LS

?

SOS1: NS NF1: NF1

Figure 1.3: Regulation of Ras/MAPK pathway by SHP2 and genes aﬀected in developmental disorders.

RTK: Receptor tyrosine kinase. SOS1: Son Of Sevenless 1. NF1: Neuroﬁbromin 1. NF1: Neuroﬁ- bromatosis type 1. NS: Noonan syndrome; LS: LEOPARD syndrome; CS: Costello Syndrome; CFCS:

Cardio-facio-cutaneous syndrome. Modiﬁed from Tartaglia et al. (2010).

In 2011, Kratz et al. revised 1900 cases of diverse RASopathies reported in the literature since 1937 and its association with cancer. They found that, indeed, there is an increased incidence of cancer, particularly in patients with Costello syndrome (11%), followed by Noonan (3.9%), CFC (3.5%) and LEOPARD (1.6%) syndromes. The cancer types

(17)

myeloid leukemia (AML), juvenile myelomonocytic leukemia (JMML) to neuroblastomas.

1.2.1 Noonan syndrome

In the late 50's the pediatrician Jacqueline Noonan described a new syndrome that had similarities in phenotype with the previously reported Turner syndrome but with associated congenital heart disease (Noonan and Nadas, 1958). The Noonan syndrome (OMIM163950) is a relatively common autosomal dominant disorder with an estimated incidence of 1 in 1000-2500 live births. The most characteristics of Noonan patients comprise dysmorphic facial features, proportionate short stature, pulmonic stenosis and hypertrophic cardiomyopathy, webbed neck, chest deformity, cryptorchidism, mental retardation and bleeding diatheses.

It took approximately 40 years to identify the genes responsible for this syndrome.

Tartaglia et al. (2001) performed a mutation screening of two families with Noonan syndrome and identiﬁed a series of substitution mutations in the genePTPN11. These missense mutations were found to be involved in switching the SHP2 protein into a constitutionally active conformation. Most of the Noonan-associated mutations are located in the exon 3, which encodes for the N-SH2 domain and in the PTP domain.

In addition toPTPN11, germline mutations inKRAS,RAF1 andSOS1 have been found to be associated with Noonan syndrome, though in a lower frequency.

Figure 1.4: Mutations associated with Noonan syndrome and leukemia. From Grossmann et al. (2010).

Then, Araki et al. (2004) generated a knock-in mouse model for the Noonan syndrome by inserting the Noonan-related mutation D61G by cre recombination. Homozygous mice for the D61G mutation died, whereas less than 50 % of heterozygous mice were viable. Here, they demonstrated that the SHP2 phosphatase activity increased, while the highest level was reached in homozygotic cells. Consistent with the phenotype observed in individuals with Noonan syndrome, heterozygotic mice also showed short stature

(18)

and, after some months, mice developed splenomegaly and myeloid expansion. They concluded that D61G+/- mice developed a myeloproliferative syndrome.

1.2.2 LEOPARD Syndrome

Leopard syndrome (LS; OMIM 151100) is a rare multisystemic disorder, mainly characterized by facial, skin and cardiac anomalies. LEOPARD is an acronym described by Gorlin et al. (1969) that resumes the major features that characterize this disorder.

By 2008, there were approximately 200 LEOPARD patients worldwide, though it is considered that there are many underdiagnosed or misdiagnosed cases (Sarkozy et al., 2008).

Mutations associated with this disorder have been identified mostly in the genePTPN11 and RAF1 (Tartaglia and Gelb, 2005; Pandit et al., 2007). Interestingly, the mutations associated with NS and LS are exclusive. Most NS mutations occur within the N-SH2 domain that results in gain-of-function with increased phosphatase activity (Keilhack et al., 2005). However, LS mutants in zebrafish are found to have dominant negative effects (Jopling et al., 2007).

1.2.3 Cardio-facio cutaneous syndrome

The CFC syndrome (OMIM115150) was ﬁrst described in the late 1980’s by Reynolds et al.

and Neri et al. Typical manifestations include congenital heart defects, characteristic facial appearance, ectodermal abnormalities and mental retardation. CFC patients carry germline mutations in four diﬀerent genes: KRAS,MEK1,MEK2 andBRAF (for review, see Roberts and Der, 2007). Approximately 75% of the patients have BRAF mutations, found to be the most frequently mutated locus in CFC patients.

Anastasaki et al. (2009, 2012) expressed a panel of BRAF and MEK alleles in zebraﬁsh embryos. Both kinase-activating and kinase-impared CFC mutants promoted similar developmental outcome during early development. There was a developmental time window in which a constant low-dose therapeutic MEK inhibition restore the normal development.

In 2011, Urosevic et al. presented a mouse model for the CFC syndrome with a germline mutation in the V600E hypomorphic allele which resembled partially phenotypical aspects observed in humans, including cardiomegaly, small dysmorphism and a reduced life span.

However, these mice developed neuroendocrine tumors, which have not been observed in CFC patients.

(19)

1.3 MUTANOM Consortium

This thesis was accomplished as part (subproject 7) of the MUTANOM Project - Sys- tems Biology of Genetic Diseases funded through the NGFN Plus Research Initiative (http://www.mutanom.org/). The overall objectives of the MUTANOM consortium were to characterize the functional consequences of somatic mutations in cancer and to develop models that predict the outcome of such genetic alterations on a molecular pathway, cellular and organism level. Initially the consortium concentrated on characterizing driver mutations selected from databases and publications (Table 1.1). Based on comple- mentary expertise in the fields of proteomics, functional genomics (available from the partner institutions Max-Planck-Institute for Molecular Genetics, Max-Delbrück-Center for Molecular Medicine, German Cancer Research Center Heidelberg and Charité), a systematic assessment of the downstream consequences of driver mutations using mass spectrometry analysis, expression profiling and phenotypic analysis was performed (Fig.

1.5).

The task of subproject 7 (Cellular Signalling Networks) was to investigate the roles of candidate genes in controlling proliferation, cellular survival and various neoplastic properties. The standard approach for testing putative oncogenes was to express the candidate gene and its mutated derivative under the control of heterologous promoters in appropriate recipient cells and to assess their impact on cellular parameters, typically associated with the transformed state such as proliferation without anchorage. Conversely, putative tumor suppressing activity of candidate genes was assayed in tumorigenic cell lines by RNA interference or forced expression of the candidate gene in tumorigenic cell lines (see Ph.D. thesis by Sha Liu, in preparation). At the molecular level, the effects of candidate genes on receptor tyrosine kinase/Ras/ MAP-kinase signal transduction and related pathways were analyzed. To begin to understand candidate gene effects at the systems level, their impact on the genetic program of cells expressing the candidate cancer or anti-cancer gene by expression profiling was assessed as well.

Table 1.1: Cancer mutations selected for analysis in the MUTANOM consortium.

Somatic mutations in the candidate genes were selected according to their number of reported cases from the COSMIC database.

Gene Gene name

Mutation Nucleotide Aminoacid APC Adenomatous polyposis coli 4348C→T R1450∗

4666insA T1556fs*6

BRAF v-raf murine sarcoma 1798GT→AA V600K

viral oncogene homolog B1 1799T→A V600E

770A→G Q257R

1399T→G & S467A 1455G→C & L485F 1495A→G & K499E

CDH1 cadherin 1, type 1, 786 794CAC

CCAGGA→T

T263fs*3

E-cadherin (epithelial) 1108G→C D370H

CDKN2A cyclin-dependent kinase inhibitor 2A 172C→T R58*

(melanoma, p16, inhibits CDK4) 238C→T R80*

(20)

Gene Gene name

Mutation (continued)

Nucleotide Aminoacid CTNNB1 catenin (cadherin-associated protein), 121A→G T41A

beta 1, 88kDa 134C→T S45F

c.110C→T S37F

EGFR epidermal growth factor receptor del2235 2249 del E746-A750

2573T→G L858R

FBXW7 F-box and WD repeat domain containing 7

1393C→T R465C

1394G→A R465H

HRAS v-Ha-ras Harvey rat sarcoma 182A→G Q61R

viral oncogene homolog 35→G G12V

350A→G K117R

IDH1 Isocitrate dehydrogenase 1 (NADP+), soluble

395G→A R132H

JAK2 Janus kinase 2 1849G→T V617F

1624-

1629delAATGAA

N542 E543del

16111616delTCA CAA

F537 K539→L

1615 1616AA→TT K539L KIT Mast/stem cell growth factor receptor

Kit

1676T→A V559D

(Proto-oncogene tyrosine-protein 2447A→T D816V kinase Kit (c-kit) (CD117 antigen) 1669 1672TGGA→G W557 K558del

1509 1510insGC CTAT

Y503 F504insAY

KRAS v-Ki-ras2 Kirsten rat sarcoma 35G→A G12D

viral oncogene homolog 35G→T G12V

101C→G P34R

458A→T D153V

467C→A F156L

MLH1 mutL homolog 1, colon cancer, 1151T→A V384D

nonpolyposis type 2 697T→C C233R

MSH6 mutS homolog 6 insC3261 F1088fs*3

3261delC F1088fs*2

NF1 neuroﬁbromin 1 1381C→T R461*

2033delC P678fs*10

NRAS neuroblastoma RAS 37G→C G13R

viral (v-ras) oncogene homolog 35G→A G12D

182A→G Q61R

181C→A Q61K

NRK Nik related kinase 1270A→T S424G

PIK3CA phosphoinositide-3-kinase, 1633G→A E545K catalytic, alpha polypeptide 3140A→G H1047R

c.1258T→C C420R

PTCH1 patched homolog 1 2975A→G E992G

PTEN phosphatase and tensin homolog 697C→T R233*

800delA K267fs*9

388C→G R130G

389G→A G132

PTPN11 protein tyrosine phosphatase, 227A→G E76G

non-receptor type 11 226G→A E76K

417G→T E139D

1403C→T T468M

RET ret proto-oncogene 2753T→C M918T

1894 1906→AGCT E632 T636→SS

1900T→C C634R

MYLK4 Myosin Light Chain Kinase Family, 232G→T A78S

(21)

Gene Gene name

Mutation (continued)

Nucleotide Aminoacid

733C→T Q245*

SMO smoothened homolog c.1210G→A V404M

1604G→T W535L

1918A→G T640A

SRC v-src sarcoma (Schmidt-Ruppin A-2) viral oncogene homolog

1591C→T Q531*

STK32B serine/threonine kinase 32B 940G→T E314*

TP53 tumor protein p53 743G→A R248Q

818G→A R273H

c.524G→A R175H

Synthesis of wildtype and mutant genes Selection of candidate mutations

Selection of gene transfer approach

HEK-TREx cells

Isogenic, dox-inducible

Selection of cell system/

Generation of stable cell lines

Effects on signaling

mRNA profiling Protein-protein

interaction

Cell mophology Cell proliferation Anchorage-independent

growth Tumor xenograft

growth Functional

assays Generation of expression clones

(Gateway® cloning system)

Tandem affinity Purification (TAP)

Co-IP Yeast-two- hybrid (Y2H)

Illumina microarrays Reverse phase protein array (RPPA)

Western blot

Western blot MUTANOM: selection of candidate genes

Figure 1.5: Experimental workﬂow of the MUTANOM consortium

(22)

1.4 Aims of this work

This thesis focuses on the characterization of mutations in PTPN11/SHP2 and BRAF genes. Both genes have been associated with the so-called ”RASopathies”, a group of developmental disorders caused by germline mutations of members of the RAS/MAPK pathway (Tartaglia and Gelb, 2005), and with somatic mutations that have been found in diﬀerent forms of leukemia. (Grossmann et al., 2010).

Although somatic mutations in SHP2 associated with cancer have been phenotypically well characterized, it is still unclear whether RASopathies-associated mutations have the potential for oncogenic transformation.

Therefore, the aim of this work was to investigate the inﬂuence of leukemogenic and RASopathies-associated mutations in SHP2 and BRAF on the cellular phenotype, proliferation and anchorage-independent growth using non-transformed cell systems.

Furthermore, to elucidate the molecular mechanisms that stimulate modifications of the cellular phenotype, protein signaling and gene transcription analysis using high- throughput methods were explored. The experimental work-flow applied for this study is shown in fig. 1.6.

(23)

Synthesis of wildtype and mutant genes Synthesis of wildtype and mutant genes

Selection of candidate PTPN11/SHP2 and BRAF mutations

Selection of gene transfer approach

208F rat fibroblasts Preneoplastic, non-transfomed HEK-TREx cells³

Isogenic, dox-inducible

Selection of cell system/

Generation of stable cell lines²

mRNA profiling Protein-protein

interaction

Cell mophology Cell proliferation Anchorage-independent

growth Tumor xenograft

growth Functional

assays Generation of expression clones

(Gateway® cloning system)

Tandem affinity Purification (TAP)

Co-IP Yeast-two- hybrid (Y2H)⁴

Illumina microarrays Reverse phase protein array (RPPA)

Western blot

Western blot MUTANOM: selection of candidate genes¹

Figure 1.6: Experimental workﬂow of this project

1Following cell lines were generated for the MUTANOM consortium as part of the selection of candidate cancer mutations for characterization: stable TAP-tagged expressing HEK-TREx cell lines were generated for TAP assay and tested for protein expression of the Ras-MAPK signaling components by western blot (each with and without Stop codon): A). K-Ras wt, G12D (35G→A), G12V (35G→T), D153V (458A→T), P34R (101C→G) and F156L (467C→A); B). SMAD4 wt, Q245∗(733C→T) and D351H (1051G→C).

2 Human breast epithelial MCF10A populations were generated by lentiviral transduction to stable express H/K-Ras wt and G12V and BRAF wt, V600E and V600K. The resulting cell lines were tested for cell morphology and proliferation, anchorage-independent growth (except for BRAF V600K), and activation of the MAPK signaling pathway.

3 To test whether the YFP-tagged TREx-HEK cell system was suitable for anchorage-independent growth assay, a soft agar assay was performed with the following cell lines (YFP-tagged expressing cells generated by Sha Liu): K-Ras wt and its corresponding mutations G12D, G12V, P34R, D153V and F156L. The TREx-HEK cell system resulted not appropriate for functional assays due to its ability to form colonies in soft agar assay without the expression of the corresponding oncogene mutation. An additional test under serum starvation conditions (0.2% FCS) showed that the HEK-TREx cells were unable to form colonies even though an oncogene such as H-Ras G12V was expressed.

4 Yeast-two-hybrid assay was performed with SHP2 wt.

(24)

2 RESULTS

2.1 Generation of an eﬃcient gene transfer approach to meet Mutanom requirements

2.1.1 Description of the selected mutations

To select the appropriatePTPN11 and BRAF mutations, a wide literature search was performed using the PubMed database. The query was focused on reported mutations without a biological characterization. Additionally, a search was performed in the Catalogue of Somatic Mutations in Cancer (COSMIC) (Forbes et al., 2008) for frequency in leukemia and solid tumors (Table 2.1).

Table 2.1: Missense mutations inPTPN11 andBRAF selected for functional studies.

Gene Substitution

Syndrome Cancer type Reference Nucleotide Aminoacid

PTPN11

124A→G T42A NS - (Tartaglia et al., 2002)

228G→C E76D NS - (Tartaglia et al., 2002)

227A→G E76G NS Colon adenocarci-

noma, JMML

(Tartaglia et al., 2003) 226G→A E76K NS JMML, AML (Tartaglia et al., 2003)

417G→T E139D NS JMML (Tartaglia et al., 2002)

844A→G I282V NS - (Tartaglia et al., 2002)

1403C→T T468M LS Rectal adenocarcinoma

(Digilio et al., 2002)

BRAF

770A→G Q257R CFC - (Niihori et al., 2006;

Rodriguez-Viciana et al., 2006)

1399T→G S467A CFC - (Rodriguez-Viciana et al.,

2006)

1455G→C L485F CFC Malignant

melanoma

(Niihori et al., 2006;

Rodriguez-Viciana et al., 2006; Gallagher et al., 2008)

1495A→G K499E CFC - (Niihori et al., 2006)

1799T→A V600E* - Malignant

melanoma, thyroid carcinoma, colon cancer

(Davies et al., 2002)

*control; NS: Noonan syndrome; NS/JMML: Noonan syndrome with juvenile myelomonocytic leukaemia;

LS: LEOPARD syndrome: lentigines, ECG conduction abnormalities, ocular hypertelorism, pulmonic stenosis, abnormal genitalia, retardation of growth, and sensorineural deafness syndrome; CFC: Cardio- facio-cutaneous syndrome; JMML: juvenile myelomonocytic leukaemia; AML: acute myeloid leukaemia.

The reference sequences, corresponding to Homo sapiens, were taken from the Concensus Coding Sequence Database (CCDS), accession numbers CCDS9163.1 forPTPN11/SHP2

(25)

sized containing a stop codon (TAA) and assembled into a vector backbone containing the ﬂanking sequences attB for posterior gateway cloning (GeneArt, Germany).

BRAF^K⁴⁹⁹Ê, SHP2^wt and its mutant derivates were assembled into the vector pMK-RQ (Kanamycin^r), BRAF^L⁴⁸⁵^F in pMA (Ampicillin^r), BRAF^S⁴⁶⁷Â and BRAF^Q²⁵⁷^R in pMS (Spectinomycin^r). BRAF^wt and BRAF^V⁶⁰⁰Ê were also synthesized by GeneArt and assembled in the entry vector pDONR221 (kindly provided by Bodo Lange, MPI, Berlin).

The mutations are located across all domains of both SHP2 and BRAF proteins as indicated in ﬁg. 2.1.

BRAF

E139D* I282V

T468M⁺

N-SH2 C-SH2 PTP

Y542 Y580 T42A

E76D E76G*⁺

E76K*

Q257R

R

RBD CR CR2 Kinase - CR3

CR1

S467A L485F K499E SHP2

Figure 2.1: SHP2 and BRAF protein domains with the localization of leukaemogenic, Noo- nan/LEOPARD and CFC mutations. N-SH2: N-terminal src-homology domain; C-SH2: C-terminal src-homology domain; PTP: Phosphotyrosine domain. Mutations marked with an asterisk (*) correspond to those reported in juvenile myelomonocytic leukaemia (JMML) and with a cross (+) reported in adenocarcinoma. N-SH2: N-terminal src-homology domain; C-SH2: C-terminal src-homology domain;

PTP: protein tyrosine phosphatase domain. CR: conserved region; CR1 corresponds to the Ras-binding domain (RBD) and a cystein-rich domain (CR); CR2: serine- and threonine-rich regulatory domain;

CR3: kinase domain.

2.1.2 Selection of the gene transfer conditions

To develop a pipeline for the evaluation of mutations by functional assays, diﬀerent mammalian cell lines were tested for overexpression by combining transfection reagents, transfection time and expression vectors.

The ﬁrst issue to assess was the suitability of the expression vector. The isogenic and doxycycline-inducible TREx-HEK cell system, which was selected by the MUTANOM consortium for functional and high-throughput assays, resulted not appropriate for functional assays due to its ability to form colonies in soft agar assay without the expression of the corresponding oncogene mutation (Sha Liu, personal communication).

An additional test under serum starvation conditions (0.2% FCS) showed that the HEK- TREx cells were unable to form colonies even though an oncogene such as H-Ras^G¹²^V was expressed.

For this reason, the following target cell lines were used for selection of the gene transfer conditions: 208F (rat ﬁbroblasts), MCF10A (human breast epithelial cell line) and

(26)

Cos7 (simian kidney fibroblast-like cells), this latter used as an easy-to-transfect control, were transiently transfected with the CMV-driven, YFP-tagged expression vector N- eYFP-amp carrying HRas^wt. The cells were treated with different transfection reagents (Lipofectamine2000, Polyethylenimine [PEI], Amaxa, Fugene and Effectene) overnight, and the transfection efficiency was monitored 48h later by fluorescence microscopy.

As expected, Cos7 cells showed a moderate to high transfection efficiency with all tested reagents. Amaxa nucleofection and Lipofectamine2000 transfection were toxic for MCF10A, whereas PEI and Effectene did not affect the cell viability, though they were not effective for the target cell lines. Fugene was the less toxic reagent but still not efficient enough to detect YFP protein expression in 208F and MCF10A. In contrast, the transfection efficiency of the empty vector, as well as the YFP-amp-Ras vector in Cos7 cells was moderate (Fig 2.2A). Since the expressed proteins were N-terminal YFP-tagged, it might be possible that the YFP-tag or the polypeptide linker interfered with the protein folding, thus affecting the protein structure as reported previously (Prescott et al., 1999). To test the protein expression of YFP-HRas cell lysates were obtained from transiently transfected cells (96-120h after transfection) and subjected to SDS-PAGE and western blot. Both Cos7 and 208F cells showed a homogeneous expression of endogenous H-Ras protein. Although YFP-H-Ras was expressed with the predicted protein size in Cos7 cells under the CMV promoter, 208F cells failed to expressed YFP-HRas (Fig.

2.2B). These results are consistent with the lack of ﬂuorescence in 208F cells compared to Cos7.

(27)

Cos7

208F

MCF10A N-YFP-Amp H-Ras^wt

(-)

208F

YFP-Ras

End. Ras

Actin

50 KDa 21 KDa

A

B ^Cos7

(-)

Figure 2.2: Transfection eﬃciency of target cell lines. Cos7, 208F and MCF10A were transfected with the expression vector N-YFP-Amp or N-YFP-H-Ras^wt. 48h after transfection, the expression of the YFP-HRas protein was monitored by ﬂuorescence microscopy (A). 96h-120h after transfection, cells were subjected to SDS-lysis. Ras protein expression was evaluated by western blot for Cos7 and 208F cells (B). (-): parental cell line, Empty: N-YFP-Amp, H-Ras^wt: N-YFP-H-Ras^wt.

Since 208F and MCF10A were selected for functional assays, a lentiviral transduction protocol was used to enhance the gene transfer eﬃciency. This approach was chosen due to the high-eﬃciency gene transfer that is required for a well-detectable and constitutive expression of the gene of interest. For this purpose, SHP2^wt and BRAF^wt were cloned into the lentiviral vector pLenti6-CMV-YFP to obtain a N-terminal YFP-tagged protein and lentiviral particles were produced (for method description, see 4.2.9 and 4.2.10).

208F and MCF10A cells were seeded in 6-well plates until they reached 70% confluency and infected with the corresponding high-titer lentiviral particles. YFP-tagged SHP2 expression was monitored by fluorescence microscopy 48-72h after infection. 208F cells overexpressing SHP2 showed no significant morphological changes and low fluorescence, that disappeared two weeks after being puromycin selected (data not shown).

Taking together, both expression vectors, eYFP-CMV-amp and pLenti6-CMV-YFP used for constitutive expression in 208F showed an early low transfection/transduction efficiency, but they were not able to produce YFP-protein expression over time. This fact might be explained by a possible promoter silencing effect, as it has been previously observed in the generation of human stable tumor cell lines, where the CMV-driven promoter showed a significant low expression efficiency compared to the human elongation factor 1-alpha (EF1α) promoter (Teschendorf et al., 2002). Furthermore, another

(28)

study demonstrated that when rats were given intramuscular injections of CMV-driven adenovirus containing the human ﬁbroblast growth factor 4 (hFGF4), a sustained decrease of hFGF4 transcription was observed as a result of the extensive methylation of CpG- and non CpG-sites of the CMV promoter (Brooks et al., 2004). Furthermore, Qin et al.

(2010) compared the most common used promoters across diﬀerent cell types and came to the conclusion that the CMV promoter performance varies considerably depending on the cell type in contrast to other promoter sequences such as the EF1α or the chicken β-actin promoter coupled with the CMV early enhancer (CAGG) promoters.

2.1.3 Comparison of diﬀerent lentiviral vectors

Due to the inefficient protein expression under the control of the CMV promoter, the CMV sequence of the pLenti6-CMV-YFP was excised and replaced by the EF1αpromoter (for details, see section 4.2.3). The correct insertion was verified by restriction digestion with AflII and PstI (fig. 2.3). This new vector was denominated pLenti6-EF1a-YFP.

Additionally, the lentiviral vector pCDH-EF1a-IRES-GFP, a bicistronic vector, also EF1α- driven and with an internal ribosomal enhanced sequence (IRES), was simultaneously tested.

CMV attB1 GOI V5

ClaI SpeI

eYFP attB2 ClaIlaII SSSpeSSS

EF1a

AflII PstI

L 1 2 3 4 5 C E 1 2 3 4 5 C E

6000 5000 4000 3000 2000 1500 1000 500

Kb

pLenti6-CMV

attB1 GOI attB2 eYFP V5

EF1a pLenti6-EF1a

Figure 2.3: EF1αpromoter cloning in pLenti6 expression vector. The CMV promoter was replaced by the EF1αpromoter obtained by PCR. After cloning, five clones (1-5) of the new vector pLenti6-EF1a- V5-eYFP were picked, DNA isolated and restriction digested with AflII or PstI. AflII corresponding bands for are 4466 bp, 3656 bp and 1906 bp. PstI restriction bands are 7145 bp, 1580 bp, 505 bp, 400 bp and 398 bp. C: pLenti6-CMV-V5-eYFP. E:pEF1-V5-HisC.

To test the transduction efficiency of both vectors, 208F and Cos7 were seeded in 6-well plates for posterior infection with pLenti6-CMV, pLenti6-EF1α or pCDH-IRES-GFP lentiviral particles. 48-72h after infection, cell morphology and eYFP/GFP expression was monitored by fluorescence microscopy. 208F cells performed substantially better with pCDH-IRES-GFP, showing an homogeneous fluorescence pattern and no decrease in GFP expression along time (fig. 2.4). Although pCDH-IRES-GFP did not contain a resistance

Functional characterization of cancer- and RASopathies-associated SHP2 and BRAF mutations